The PI column is used to distribute the data across the AMPS. Same data values with lend on same AMPs, which will increase the skew in tables is distribution is more on one AMP then the others.
